Introduction

Caenorhabditis elegans Maupas, 1900 has been extensively used to study the toxic effect of pollutants, drugs, and environmental samples1. One clear benefit is the exposure of an intact animal with functioning digestive, endocrine, neuromuscular, reproductive, and sensory systems2, i.e. phenotypes that are biologically relevant3. This species is also small in size, easy to culture, and can even be maintained axenically2,4. Furthermore, studies have shown that C. elegans bioassays can be used to predict mammalian development at a fraction of the cost of traditional animal testing2,5. These qualities and benefits complement C. elegans as a model organism for toxicity testing, as well as its use in high-throughput assessment protocols3, as has been developed for drugs6 and pollutants of environmental concern7.

Commonly used toxicity endpoints include feeding, fertility, growth, movement, reproduction, and survival of C. elegans1,8. Respiration of C. elegans has also been used, although infrequently, to study the effect of toxicant exposure9,10,11. This endpoint has been shown to serve as an effective measure of toxicity of especially metal pollution in, among others, microbes12, daphnia13 and earthworms14. Therefore, with the development of state-of-the-art, high-throughput respirometers, such as the Seahorse XFe96 Extracellular Flux Analyzer, the relevance and applicability of C. elegans oxygen consumption rate (OCR) measurements have greatly improved. This has led to the development of acute response protocols that measure C. elegans OCR before and after the injection of pre-loaded compounds, typically oligomycin, FCCP, rotenone and antimycin A. Such compounds facilitate the measurement of mitochondrial respiratory chain functionality in organisms by determining, for example, ATP production, proton leak, maximal respiration, spare respiratory capacity and non-mitochondrial respiration15,16,17.

However, with such short exposure periods and the lack of food, these OCR protocols have limited environmental relevance. Longer exposure periods would allow the measurement of a chronic response, which is facilitated by the short life cycle (approximately 3 days at 20 °C) of C. elegans3,18. Furthermore, by adding a food source, nematode respiration would be sustained16 and toxicant bioaccessibility increased as feeding nematodes will ingest dissolved and bacterial-bound substrates19. Therefore, the aim of this research was to develop an environmentally relevant C. elegans OCR protocol for sublethal toxicity testing by utilizing the high-throughput capabilities of the Seahorse XFe96 Extracellular Flux Analyzer.

Results and discussion

Food density

A food density bioassay was performed in order to determine the minimum amount of food (Escherichia coli OP50) required to ensure uninhibited nematode development during the present study’s experimental procedures. The relationship between food density and nematode length is illustrated in Fig. 1. Nematode length (or rather growth) was clearly inhibited by decreased food availability, a well-studied response often used to investigate the effect of dietary restrictions on C. elegans development20,21. Therefore, using a segmented regression model, the minimum density of food that allowed unrestricted nematode development was determined. This model indicated that a plateau for nematode growth was reached at a food (E. coli) OD of 0.59 (595 nm). Subsequently, an OD of 0.6 was used for the experimental bioassays.

Figure 1
figure 1

Food density bioassay. Average length and total oxygen consumption rate (OCR) of 50 larval stage one Caenorhabditis elegans nematodes are considered at different optical densities (at 595 nm) of its bacterial food source (Escherichia coli OP50) after 48 h incubation at 20 °C.

Full size image

Furthermore, the relationship between C. elegans length and OCR was investigated (Fig. 1) and visualised as a non-linear exponential growth curve (Fig. 2a). In order to infer statistical meaning from this relationship, a linear regression model (Fig. 2b) was applied. The slope of the log10 transformed linear model (\(Y=0.003761X-0.1247\)) differed significantly (P = 0.0002) from zero with nematode length explaining 98% of the variation in OCR of C. elegans. The 95% confidence bands, as illustrated in Fig. 2b, indicated a low degree of uncertainty. Previous studies have reported on the change in OCR as a function of C. elegans larval development and/or adult aging16,22,23,24,25,26. However, these observations were mainly made per life stage or for L4 and adult nematodes. It should be noted that the reason for the slight decrease in OCR at a food density of 0.7 OD (at 595 nm) (Fig. 1) remains unknown, however, it is possible, although unconfirmed, that high densities of E. coli cells can impair OCR measurement. Nonetheless, the findings presented here clearly show that following incubation, C. elegans OCR correlate to its growth stage as influenced by food availability.

Figure 2
figure 2

Oxygen consumption rate (OCR) of Caenorhabditis elegans. The OCR of Caenorhabditis elegans (of 50 larval stage one nematodes) is considered against the average length of such specimens as an (a) exponential (non-linear) growth curve and (b) log10 transformed linear model (Y = 0.003761X − 0.1247). The slope of the linear model differed significantly (P = 0.0002) from zero with nematode length explaining 98% of the OCR variation.

Full size image

Oxygen consumption rate inhibition due to toxicant exposure

For each toxicant bioassay the percentage OCR inhibition of C. elegans per exposure concentration was measured against a negative control (M9 medium). Following, concentration-response curves for Benzylcetyldimethylammonium chloride monohydrate (BAC-C16) (R2 = 0.93; P = 0.001) (Fig. 3) and cadmium (Cd) (R2 = 0.98; P = 0.001) (Fig. 4) were created and used to derive ECx values (Table 1) at the 10, 20, and 50% inhibition levels.

Figure 3
figure 3

Oxygen consumption rate (OCR) inhibition following benzylcetyldimethylammonium chloride monohydrate (BAC-C16) exposure. Concentration-response curve of Caenorhabditis elegans oxygen consumption rate (OCR) inhibition following exposure to BAC-C16. The R2 value was calculated as 0.93 (P = 0.001) and 95% confidence bands are indicated as dotted lines.

Full size image
Figure 4
figure 4

Oxygen consumption rate (OCR) inhibition following cadmium (Cd) exposure. Concentration-response curve of Caenorhabditis elegans oxygen consumption rate (OCR) inhibition following exposure to Cd. The R2 value was calculated as 0.98 (P = 0.001) and 95% confidence bands are indicated as dotted lines.

Full size image
Table 1 Effective concentration (EC) values representing 10, 20 and 50% of Caenorhabditis elegans oxygen consumption rate (OCR) and growth inhibition.
Full size table

The BAC-C16 and Cd concentrations at which 50% OCR inhibition of C. elegans occurred were calculated as 8.94 mg/L and 0.73 mg/L (Table 1), respectively. Benzylcetyldimethylammonium chloride monohydrate is routinely used as a positive control for C. elegans toxicity assays27,28, while also serving as the positive control in standardised toxicity testing29. Although C. elegans OCR inhibition following exposure to BAC-C16 has never been investigated, Schouest et al.11 used fluorescence oxygen sensing and recorded significant OCR reduction following 24 h exposure of C. elegans adults to Cd, as well as other toxicants (e.g. zinc and rotenone). The latter authors calculated an EC50 value (for OCR inhibition) following exposure to Cd of 60.85 µM (6.84 mg/L). This is substantially higher than the EC50 value reported for this study (0.73 mg/L). However, in the present study longer exposure periods of 48 h (vs 24 h) and larval stage one (vs adult) nematodes were used. It is known that longer exposure periods can lead to greater toxicity30, while larval stages are typically more sensitive to metal exposure than adults31.

Also worth considering is the sensitivity of OCR compared to reproduction, since the latter is regarded as one of the most sensitive C. elegans endpoints of toxicity19,29. Comparing our results to the findings of other studies, it seems that OCR might be slightly less sensitive than reproduction (EC50 of BAC-C16: 7.5 mg/L; EC50 of Cd: 0.21 mg/L)19,27. However, these direct comparisons should be made with caution since different exposure periods were used for deriving ECx values for OCR (48 h) and reproduction (96 h).

Growth inhibition due to toxicant exposure

Caenorhabditis elegans growth inhibition was also investigated for two reasons: (1) a strong relationship, as evidenced in the food density bioassay, existed between C. elegans OCR and length and (2) C. elegans growth inhibition is routinely used as an endpoint of toxicity and can therefore be used to evaluate the sensitivity of C. elegans OCR inhibition as an endpoint of toxicity. Growth inhibition concentration-response curves for BAC-C16 (R2 = 0.97, P = 0.001) and Cd (R2 = 0.95, P = 0.001) are illustrated on Figs. 5 and 6, respectively.

Figure 5
figure 5

Growth inhibition following benzylcetyldimethylammonium chloride monohydrate (BAC-C16) exposure. Concentration-response curve of Caenorhabditis elegans growth inhibition following exposure to BAC-C16. The R2 value was calculated as 0.97 (P = 0.001) and 95% confidence bands are indicated as dotted lines.

Full size image
Figure 6
figure 6

Growth inhibition following cadmium (Cd) exposure. Concentration-response curve of Caenorhabditis elegans growth inhibition following exposure to Cd. The R2 value was calculated as 0.95 (P = 0.001) and 95% confidence bands are indicated as dotted lines.

Full size image

The BAC-C16 concentration at which 50% growth inhibition of C. elegans occurred was calculated as 9.47 mg/L (Table 1). Although EC values of C. elegans OCR inhibition for BAC-C16 have not previously been reported, ISO1087229 states that the EC50 value for BAC-C16 growth inhibition typically ranges between 8 and 22 mg/L. Most studies have reported EC50 values of approximately 15 mg/L27,28, however, Schertzinger et al.32 reported EC50 values for two separate tests of 9.1 and 10.8 mg/L, respectively. The latter values are therefore similar to results reported in this study.

According to Hanna et al.28 growth inhibition by BAC-C16 is substantially influenced by food density, with lower densities presenting higher inhibition rates. However, this was not viewed as a concern during the present study as the minimum required amount of food for C. elegans was used at a constant density in control and exposure wells of toxicant bioassays.

The Cd concentration at which 50% growth inhibition of C. elegans occurred was calculated as 0.86 mg/L (Table 1). As with BAC-C16, Cd toxicity has been linked to food density, with increasing toxicity as food density decreases19. The latter authors attributed this to a decrease in the bioavailability of freely dissolved Cd at high bacterial densities. Traunspurger et al.33 reported a lowest observed effect concentration (LOEC; 72 h exposure) of 0.14 mg/L for the growth of C. elegans. This can be compared to the present study’s EC10 and EC20 values of 0.43 and 0.56 mg/L, respectively, following 48 h exposure. Van Kessel et al.34 on the other hand, showed a substantially higher LOEC (11.2 mg/L) for C. elegans after 48 h exposure to Cd. However, this was in the absence of food, which results in highly reduced Cd bioaccessibility. The presence of bacteria (food source) stimulates C. elegans pharyngeal pumping, which promotes the uptake of dissolved and bacterial-bound Cd19.

Comparing OCR and growth inhibition

Similar ECx values for OCR and growth inhibition of C. elegans were evidenced (Table 1). Furthermore, the correlations between OCR and growth inhibition for BAC-C16 and Cd were strong [R2 = 0.93 (P = 0.001) and R2 = 0.91 (P = 0.001), respectively] and thus support the relationship evidenced between nematode length and OCR in the food density bioassay.

Lastly, this protocol and the use of C. elegans as a test organism can be further studied by, for example, investigating the sensitivity of C. elegans organismal vs. cellular respiration15,16,17. Considering that for the tested toxicants, OCR was equally sensitive to growth (an accepted and routinely used endpoint of toxicity1,29) we have clearly demonstrated that OCR inhibition can be used as an alternative method that facilitate high-throughput and rapid measurements.

Quality control

A number of steps were taken to maximise the confidence in the generated results. Firstly, for accurate OCR measurements, oxygen consumption from other organisms than the test organism should be excluded. Since E. coli, serving as food source for the C. elegans, is an aerobic bacterium, the bacterial cultures had to be inactivated before use [see Stock preparation of the food source (Escherichia coli)]. Additionally, Penicillin-Streptomycin (Pen-Strep) was added to each well (see Experimental procedure) to ensure the inactivation of the E. coli food source and prevent bacterial contamination. Bacterial inactivation was checked by measuring the OCR of four extra wells with inactivated bacteria, but without nematodes (see Bioassay plate layout). In case of any oxygen consumption in these “E. coli” controls, the whole bioassay plate was discarded.

Secondly, temperature drifts occur during the operation of the Seahorse respirometer (see Temperature requirements). However, using a Seahorse respirometer, Koopman et al.16 studied the OCR of L4 and adult C. elegans and found no significant (P > 0.05) difference in OCR between 20 °C and 25 °C. Also in this study, OCR data measured with the Seahorse respirometer at 20 °C and 24.5 °C were not significantly different (P > 0.05), indicating that temperature shifts in this range have only a negligible influence on the outcome of the bioassays.

Lastly, all wells of the bioassay plate should be sufficiently oxygenated during OCR measurement in order to avoid inaccurate readings. During this study, the unprocessed oxygen level data (in mmHg) was checked (following OCR measurement) in order to confirm that anoxic conditions were never reached.

Final considerations

The results presented in this study provide support for the use of C. elegans OCR inhibition as a functional and ecologically relevant endpoint of toxicity. Although different methods and instruments can be utilized, the high-throughput capability of modern respirometers facilitate simultaneous and rapid toxicity measurement of substances at various concentrations. Also, with further advancement and optimization of respirometers (e.g. better temperature management) it is likely that preparation and operating times will be reduced and measurement accuracy increased. Lastly, the authors recommend to explore more appropriate bacterial inactivation methods (e.g. without the addition of chemicals), or even the use of anaerobic bacteria as a food source.

Methods

Cultures and reagents

Cultures of C. elegans N2 and E. coli OP50 (food source) were obtained from the Caenorhabditis Genetics Centre (https://cbs.umn.edu/cgc/home). Stock solutions of sterile M9 medium (buffer) and cholesterol were prepared, as well as cultures of C. elegans reared, following ISO1087229. Penicillin-Streptomycin (Gibco 100X) was obtained from Thermo Fisher Scientific while the remaining reagents used in this study were obtained from Sigma-Aldrich®.

Stock preparation of the food source (Escherichia coli)

An important step in the preparation of food stocks is the culturing, inactivation, and density adjustment of E. coli before the commencement of a bioassay. Escherichia coli was cultured, washed, and pelleted according to ISO1087229. The pellet was re-suspended in M9 medium after which an aliquot was diluted (1 → 10) and the optical density (OD) measured at 595 nm35 using a Pharo 300 Spectroquant (spectrophotometer). Finally, the density of the E. coli suspension was adjusted to an OD of 3 (595 nm). Thereafter, 5 mL aliquots of food stocks were transferred to 15 mL conical centrifuge tubes and heat inactivated (30 min at 65 °C) using a water bath36. The stocks were stored at −80 °C for a maximum of two weeks.

Synchronization of Caenorhabditis elegans

A schematic overview of the experimental procedures, as well as the appropriate timeline, is provided in Fig. 7. On day one, C. elegans eggs were extracted from culture plates using the sodium hypochlorite (bleaching) method, followed by three wash cycles to remove any residual chemicals15. Subsequently, synchronized larval stage one (L1) nematodes were obtained after overnight (12–20 h) incubation in 50 mL sterile conical centrifuge tubes at 20 °C on an orbital shaker (100 rpm)15. This procedure also kills and dissolves E. coli, which could otherwise influence the outcome of the bioassay. It should be noted that prior to bleaching, the culture plates were studied using a Nikon SMZ1000 stereo microscope (100x magnification) to confirm the presence of eggs and gravid females.

Figure 7
figure 7

Experimental setup. Schematic overview of the experimental procedure and associated timeline for culturing Caenorhabditis elegans and its bacterial food source (Escherichia coli OP 50). The timeline for preparing and incubating the bioassay plate, as well as for C. elegans oxygen consumption rate (OCR) measurement using the Seahorse XFe96 Extracellular Flux Analyzer, is also provided.

Full size image

Number of nematodes

An important consideration is the number of nematodes per well to be used for OCR measurements. In order to effect a broad OCR range, 50 L1s per well were used. Koopman et al.16 reported this to be the largest number of larval stage four (L4) (expected life stage after incubation) nematodes to be used per well without inducing anoxic conditions.

Food density and nematode development bioassay

Since food availability can have a substantial impact on C. elegans development36, a food density (0.1–0.7 OD, 595 nm) bioassay (see Bioassay plate layout) was performed (as described in Experimental procedure), with 50 L1s per well. The OCR of C. elegans was measured as detailed in Seahorse respirometer setup and oxygen consumption rate measurement.

Toxicant stock solutions

In order to test the viability of OCR inhibition of C. elegans as a sublethal endpoint of toxicity, benzylcetyldimethylammonium chloride monohydrate (BAC-C16) and cadmium (Cd) were selected as toxicants. Benzylcetyldimethylammonium chloride monohydrate is routinely used as a positive control for C. elegans growth inhibition27,28,29, while Cd is regarded as an environmentally relevant, non-essential metal19,37,38.

The exposure solutions for BAC-C16 (made up in M9 medium) had the following concentrations: 2, 4, 6, 8, 10, 12, 16, 20, and 24 mg/L. For Cd, the exposure concentrations in M9 medium were 0.25, 0.5, 0.75, 1, 0.25, 1.5, 1.75, and 2 mg/L. For both assays (designed, performed, and measured as described in the following sections) the negative control consisted of M9 medium.

Bioassay plate layout

The bioassay was carried out in 96-well culture plates, which will from hereon be referred to as ‘bioassay plates’. The bioassay plate layout (Fig. 7) was designed to allow for the maximum number of exposure concentrations with sufficient replication. The four corners (A1, H1, A12, and H12) represented the baseline wells, which the Seahorse respirometer uses for background correction of zero oxygen consumption. Four wells (B1, G1, B12, and G12) labelled “Ecoli” were reserved for E. coli food stocks (0.6 OD, 595 nm) containing a biocontrol solution (Pen-Strep and cholesterol), brought to a final volume of 200 µL with M9 medium. These wells were included to ensure that zero E. coli oxygen consumption, which could substantially impact OCR measurements, occurred (also see Quality Control).

Wells C1 – F1 and C12 – F12 were assigned for future use as a positive control for which BAC-C16 is recommended at a concentration of 8.94 mg/L (EC50 value of OCR inhibition as determined during this study). For the positive control the oxygen consumption rate inhibition of C. elegans (compared to the negative control), should range between 20% and 80%29. Column 2 was represented by the negative control, while columns 3 to 11 were used for the measurement of nine treatments with eight replicates each.

Experimental procedure

Working in a sterile environment, the following preparation steps were executed on day two (Fig. 7):

  1. 1.

    A tube of E. coli food stock was allowed to thaw and reach room temperature.

  2. 2.

    Synchronized L1 nematodes were pelleted (2 min at 2000x g) and the supernatant discarded in order to remove residual material. Next, the pellet was re-suspended in 5 mL M9 medium using a magnetic stirrer. It should be noted that studying L1 nematodes using a Nikon Eclipse 50i light microscope (1000x magnification) after being subjected to stirring revealed no physical damage. Similarly, Van Aardt et al.39 reported that stirring speed had no effect on the OCR of Meloidogyne incognita (Kofoid & White, 1919) Chitwood, 1949 second-stage juveniles.

  3. 3.

    Suspended nematodes were transferred, in 10 µL aliquots, to a microscope slide and counted using a Nikon SMZ1000 stereo microscope (40-100x magnification). The average number of nematodes in 10 replicates were calculated per 1 µL. Thereafter, the concentration of nematodes in suspension was adjusted to 50 individuals per 35 µL.

  4. 4.

    Stock solutions of the studied toxicants (BAC-C16 and Cd) were prepared at twice the concentration of the highest exposure concentration as stock solutions were diluted 1:2 with food, biocontrol solution (see below), and nematode suspensions (Fig. 7). By taking into account the final volume (200 µL) of each well, the required volume of toxicant stock was calculated for each exposure concentration using the following equation:

    $${V}_{1}=\frac{{C}_{2}{V}_{2}}{{C}_{1}}$$

    where V1 represents the volume (unit: µL) of toxicant stock, C2 the final exposure concentration (unit: mg/L), V2 the final well volume of 200 µL, and C1 the toxicant stock concentration (unit: mg/L).

  5. 5.

    Biocontrol solution: A solution of cholesterol stock (14 µL), Pen-Strep (200 µL), and M9 medium (5 mL) was prepared. Cholesterol is necessary for the development of C. elegans40, while Pen-Strep ensured the inactivation of the E. coli food source and prevented bacterial contamination35. Pen-Strep is frequently used in the cultivation of C. elegans41,42 and was not expected to significantly influence the results. It is also important to consider that Pen-Strep were added to both control and exposure wells, which therefore renders it potential effect (if any) on the final results negligible.

    Immediately following these preparation steps, the bioassay plate (Fig. 7) was loaded in the following sequence:

  6. 6.

    The required volume of M9 medium was added to ensure a final volume per well of 200 µL.

  7. 7.

    The calculated volume of toxicant stock per exposure concentration was loaded.

  8. 8.

    E. coli food stocks were briefly vortexed at room temperature after which 40 µL was added to all bioassay wells (excluding ‘Baseline’ wells). This ensured a final OD of 0.6 (595 nm).

  9. 9.

    Finally, 25 µL of the biocontrol solution were added to all bioassay wells (excluding ‘Ecoli’ wells).

  10. 10.

    The bioassay plate was placed on an orbital shaker (100 rpm) for 15 min in order to ensure sufficient mixture of reagents.

  11. 11.

    Lastly, 35 µL of the nematode solution was added to each control and exposure well. No nematodes were added to the ‘baseline’ or ‘Ecoli’ wells. Using a Nikon TS100 inverted microscope (40-100x magnification) each well was checked to ensure correct loading of reagents and nematodes.

The bioassay plate was sealed with parafilm and incubated for 48 h at 20 °C. This incubation period was chosen to effect high OCR readings without risk of oxygen depletion during measurement16.

Additional considerations for the execution of this protocol follows:

  • A final volume of 200 µL was assayed in order to simplify the calculation of the concentration and volume of solutions. However, according to the manufacturer (Agilent Technologies, Santa Clara, United States) the final well volume can range between 150 µL and 275 µL43. Therefore, the protocol can be adjusted accordingly if lower or higher well volumes are required.

Seahorse respirometer setup and oxygen consumption rate measurement

Cartridge hydration

The Seahorse respirometer makes use of optic fibre bundles, which insert into solid state sensor probes (containing polymer embedded fluorophores) and emit light to excite the fluorophores. These optic fibres then measure the change in fluorophore emission resulting from the change in oxygen concentration and thus serves as an indirect measurement of oxygen consumption. On day three (Fig. 7), the Seahorse respirometer cartridge, which houses the probes, was hydrated by adding 200 µL XF calibrant to each well followed by overnight incubation at 37 °C.

Temperature requirements

The Seahorse respirometer was designed for cell OCR measurement at 37 °C, contrary to C. elegans’ typical culture temperature range of between 16 and 25 °C18. While the instrument is capable of regulating the temperature in this range, it requires a room temperature of 4 °C, which was logistically not possible. Therefore, the room was cooled to the lowest possible temperature of 16 °C. The respirometer’s temperature was set to 24 °C and the internal heater switched off. It should be noted that the Seahorse respirometer generates heat during operation and was therefore only powered on directly before use. On day four (Fig. 7), the Seahorse cartridge was removed from the incubator two hours prior to use and left to cool and reach ambient temperature (16 °C).

Seahorse settings

Prior to OCR measurement, the following Seahorse respirometer protocol was programmed (Fig. 7):

  1. 1.

    Two min mixing: This involves the raising and lowering of the cartridge in order to replenish the oxygen levels within each well.

  2. 2.

    Five min waiting: The cartridge remains stationary in the ‘raised’ position to allow the nematodes to settle.

  3. 3.

    Three min measuring: The cartridge is lowered and a microchamber (of 3 µL) is created at the bottom of each well in which nematode oxygen consumption is measured. The decrease in oxygen is converted to a single OCR value per well.

This represents one measurement cycle, which was repeated eight times. After programming and 60 min prior to the OCR test, the Seahorse XF cartridge was inserted into the Seahorse respirometer for calibration. The calibration step is standard in every Seahorse analysis and cannot be unselected. During this calibration of the sensors, the instrument reads the coefficients of the sensor cartridge and bioassay plate to ensure accurate data acquisition. Once calibrated, the bioassay plate containing the nematodes was inserted into the Seahorse respirometer after which it underwent an equilibration period (12 min) during which temperature stability across the plate is ensured, followed by the above detailed OCR protocol. Oxygen levels during and after OCR measurement were checked to ensure that anoxic conditions were not induced. The Agilent Seahorse Wave 2.4 software package was used for exporting OCR data.

Upon completion of OCR measurement, 100 µL Bengal Rose (used here as a nematode staining agent) was added to each well and the bioassay plate heat inactivated (10 min at 80 °C)29. Bioassay plates were stored, for a maximum of 7 days, at 4 °C. Nematode length was measured and growth calculated as described in ISO1087229.

Statistical analyses

Food density bioassay

The average OCR (of the eight measurement cycles) and length of C. elegans were calculated per well and graphically illustrated, at different food densities, using GraphPad Prism 7 software package. Thereafter, the density of food required to allow maximum nematode development was calculated using a segmented regression model. Briefly, the growing phase of the curve was fitted with a quadratic model. The plateau, in turn, was fitted using a constant representing a line running parallel to the density food axis defining the maximum nematode development. The plateau point was determined under a condition of continuity and smoothness as defined in the supplementary material44. This analysis was performed using SAS/STAT software package 9.4.

The relationship between OCR and nematode length was explained by an exponential growth (non-linear) regression model. In order to further study this relationship, the dependant variable (OCR) was log10 transformed and a linear regression model fitted. The 95% confidence limits were also calculated. These graphs were created and analyses performed using GraphPad Prism 7 software package.

Toxicant concentration-response bioassays

The average nematode OCR and growth of the eight measurement cycles were calculated per exposure concentration for the BAC-C16 and Cd concentration-response bioassays. Using ToxRat Professional 3 software package, the percentage decrease per exposure concentration was calculated against the negative control. Thereafter, the Probit analysis using the linear maximum likelihood regression algorithm was performed. The Chi-squared test was used to indicate the goodness-of-fit of the regression line. Furthermore, the effective concentrations (EC10, EC20, and EC50) for OCR and growth inhibition of C. elegans were calculated, while 95% confidence limits were based on Fieller’s Theorem. Lastly, in order to study the relationship between OCR and growth inhibition of C. elegans, the data were tested for normality using the D’Agostino & Pearson omnibus normality test45. The data presented a normal distribution and therefore the Pearson correlation coefficient test was performed.

Oxygen consumption rate response to temperature fluctuations

In order to determine whether temperature had a significant effect on the OCR of C. elegans, the negative control data (of eight measurements) of the two toxicant bioassays were used. Temperatures during both assays ranged between 20 °C and 24.5 °C. Firstly, the bioassay data were tested for normality using the D’Agostino & Pearson omnibus normality test45. Thereafter, the significance between the measurement means were tested using a one-way analysis of variance (ANOVA) (parametric data) or Kruskal-Wallis test (non-parametric data) with Tukey’s test and Dunn’s test applied, respectively, for multiple comparisons. These analyses were performed using GraphPad Prism 7 software package.